Kwonjoon Lee

Papers

Total Citations

H-Index

About

Kwonjoon Lee is a rising researcher at the forefront of embodied AI and multi-agent systems, with a focus on bridging perception, prediction, and planning. His work tackles two critical frontiers: spatial action localization and heterogeneous multi-robot coordination. In his highly cited 2023 paper, "AdamsFormer for Spatial Action Localization in the Future" (7 citations), Lee introduced a novel task and transformer-based architecture that predicts *where* human actions will occur in future video frames—a capability essential for safe human-robot collaboration. More recently, his 2025 work, "Generalized Mission Planning for Heterogeneous Multi-Robot Teams via LLM-Constructed Hierarchical Trees" (4 citations), presents a groundbreaking approach that leverages large language models to decompose complex missions into subtasks, assigning them to robots based on their unique capabilities. This work offers a scalable, LLM-driven solution to a long-standing challenge in robotics. Though early in his career, Lee’s contributions are already shaping how autonomous systems anticipate human behavior and coordinate diverse robot teams, marking him as a promising voice in next-generation AI and robotics.